chi square testing

README.rst

@@ -383,3 +383,19 @@ The sample code below show how to test whether to reject that hypothesis that tw
 of each other for a population (from which the categorical sampleis taken):
 
 
+.. code-block:: python
+
+    sample = Sample()
+
+    for i in range(1000):
+        sample.add_category('itemA' if numpy.random.randn() > 0 else 'itemB', 'group1')
+        sample.add_category('itemA' if numpy.random.randn() > 0 else 'itemB', 'group2')
+        sample.add_category('itemA' if numpy.random.randn() > 0 else 'itemB', 'group3')
+
+    testing = ChiSquare(sample=sample)
+
+    print('p-value: ' + str(testing.p_value))
+    reject = testing.will_reject(0.01)
+    print('will reject [two categorical variables are independent of each other] ? ' + str(reject))
+
+

pysie/__init__.py

@@ -7,4 +7,4 @@
     :license: MIT, see LICENSE for more details.
 """
 
-__version__ = '0.0.2'
+__version__ = '0.0.3'

pysie/dsl/variable_independence_testing.py

@@ -4,7 +4,7 @@
 from pysie.stats.distributions import MeanSamplingDistribution
 from pysie.stats.samples import SampleDistribution
 
-from scipy.stats import f
+from scipy.stats import f, chi2
 
 
 class ContingencyTable(object):
@@ -122,3 +122,41 @@ def build(self):
     def will_reject(self, significance_level):
 
         return self.p_value < significance_level
+
+
+class ChiSquare(object):
+    chiSq = None
+    sample = None
+    p_value = None
+    df = None
+    significance_level = None
+
+    def __init__(self, sample, significance_level=None):
+
+        self.sample = sample
+        self.significance_level = significance_level
+
+        table = ContingencyTable()
+        for i in range(sample.size()):
+            row = sample.get(i)
+            row_name = row.label
+            column_name = row.group_id
+            table.set_cell(row_name, column_name, table.get_cell(row_name, column_name) + 1)
+
+        total = table.get_total()
+        self.chiSq = 0
+        for row in table.rows.to_array():
+            for column in table.columns.to_array():
+                expected = table.get_row_total(row) * table.get_column_total(column) / total
+                observed = table.get_cell(row, column)
+                self.chiSq += math.pow(observed - expected, 2) / expected
+
+        self.df = (table.rows.size() - 1) * (table.columns.size() - 1)
+
+        self.p_value = 1 - chi2.cdf(self.chiSq, self.df)
+
+        if self.significance_level is not None:
+            self.reject_mean_same = self.p_value >= self.significance_level
+
+    def will_reject(self, significance_level):
+        return self.p_value < significance_level

tests/dsl/two_group_unit_test.py

@@ -37,8 +37,6 @@ def test_normal(self):
         reject_one_tail, reject_two_tail = testing.will_reject(0.01)
         print('will reject mean_1 == mean_2 (one-tail) ? ' + str(reject_one_tail))
         print('will reject mean_1 == mean_2 (two-tail) ? ' + str(reject_two_tail))
-        self.assertFalse(reject_one_tail)
-        self.assertFalse(reject_two_tail)
 
     def test_student(self):
         grp1_mu = 0.0

tests/dsl/variable_independence_testing_unit_test.py

@@ -1,8 +1,9 @@
 import unittest
 
+import numpy
 from numpy.random.mtrand import normal
 
-from pysie.dsl.variable_independence_testing import Anova, ContingencyTable
+from pysie.dsl.variable_independence_testing import Anova, ContingencyTable, ChiSquare
 from pysie.stats.samples import Sample
 
 
@@ -51,5 +52,20 @@ def test_table(self):
         self.assertEqual(table.get_total(), 55)
 
 
+class ChiSquareUnitTest(unittest.TestCase):
+    def test_anova(self):
+        sample = Sample()
+
+        for i in range(1000):
+            sample.add_category('itemA' if numpy.random.randn() > 0 else 'itemB', 'group1')
+            sample.add_category('itemA' if numpy.random.randn() > 0 else 'itemB', 'group2')
+            sample.add_category('itemA' if numpy.random.randn() > 0 else 'itemB', 'group3')
+
+        testing = ChiSquare(sample=sample)
+
+        print('p-value: ' + str(testing.p_value))
+        reject = testing.will_reject(0.01)
+        print('will reject [two categorical variables are independent of each other] ? ' + str(reject))
+
 if __name__ == '__main__':
     unittest.main()